In computer games, tutoring systems are used for two purposes: (1) to introduce a human player to the mechanics of a game, and (2) to ensure that the computer plays the game at a level of playing strength that is appropriate for the skills of a novice human player. Regarding the second purpose, the issue is not to produce occasionally a weak move (i.e., a give-away move) so that the human player can win, but rather to produce not-so-strong moves under the proviso that, on a balance of probabilities, they should go unnoticed. This paper focuses on using adaptive game AI to implement a tutoring system for commercial games.1 We depart from the novel learning technique 'dynamic scripting' and add three straightforward enhancements to achieve an 'even game', viz. high-fitness penalising, weight clipping, and top culling. Experimental results indicate that top culling is particularly successful in creating an even game. Hence, our conclusion is that dynamic scripting with top culling can implement a successful tutoring system for commercial games. © IFIP International Federation for Information Processing 2005.
CITATION STYLE
Spronck, P., & Van Herik, J. D. (2005). A tutoring system for commercial games. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 3711 LNCS, pp. 389–400). https://doi.org/10.1007/11558651_38
Mendeley helps you to discover research relevant for your work.